Integrated land use-transport models are critical tools to assess the long-term implications of emerging transportation technologies like on-demand mobility and autonomous vehicles. We use econometric behavioral models in an agent-based microsimulation framework (SimMobility Long-Term) to model the potential impacts of an AV-friendly car-lite policy implemented in a particular neighborhood. Residential relocation and vehicle availability decisions are jointly considered in a sequential simulation approach for modeling changes in the housing-mobility bundle. Different market responses to the policy are designed through various scenarios, and compared to a baseline where the policy is never implemented. A comparatively vehicle-free study area is chosen to obtain conservative estimates of policy impacts. We find that households do find the study area more attractive. Additionally, in-movers are richer and more prone to relinquish private vehicles than out-movers. However, the net effect on vehicle ownership is less than anticipated because higher-income house-holds remain more likely to own vehicles.